MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Are there infinite sets $E\subset\mathbb{Z}$ such that any $f\in l^2(\mathbb{Z})$ with support on $E$ comes from the Fourier transform of a continuous function on $\mathbb{T}$ ? If yes, is there a characterization of these sets?

If we change the group from $\mathbb{Z}$ to the free group with n generators $\mathbb{F_n}$, $n\geq 2$,then an infinite set $E$ with the property that $x_{i_1}x_{i_2}^{-1}x_{i_3}x_{i_4}^{-1}\ldots x_{i_{2k-1}}x_{i_{2k}}^{-1}\neq 1,\ \forall k\in\mathbb{N},\ \forall x_{i_1}\neq x_{i_2},\ x_{i_2}\neq x_{i_3},\ldots x_{i_{2k-1}}\neq x_{2k}$ (studied by Leinert, Bozejko) will satisfy the following:

Any $f\in l^2(\mathbb{F_n})$ with support on $E$ will be in the reduced $C^*$ algebra $C^*_r(\mathbb{F_n})$, and moreover $$||f||_\infty\leq 2||f||_2$$

[Edit:] By $||\cdot||_\infty$ I mean the operator norm of the corresponding convolutor.

share|cite|improve this question
One characterization of an amenable group $\Gamma$ is that $\| \lambda(f) \|_\infty = \| f \|_1$ for any non-negative valued function $f \in \ell^1\Gamma$. So no amenable groups will have such an infinite subset. Which then begs the question: What about non-amenable groups which do not contain a non-cyclic free group? – Jesse Peterson Nov 13 '13 at 1:32

It depends on what you mean by "comes from the Fourier transform of a continuous function."

1) If you mean "Any $L^2$ function with Fourier coefficients support on $E$ must be continuous," then a set $E$ will have this property if and only if it is finite.

The fact that finite sets have this property is obvious. Conversely, assume that $E$ is infinite and let $\{a_n\}_{n=1}^{\infty}$ be an enumeration of your set. The function $f(x) = \sum_{n=1}^{\infty} \frac{1}{n} e(a_n x) $ will be in $L^2(\mathbb{T})$ but will have a singularity at $0$ and thus won't be continuous on $\mathbb{T}$. [This isn't quite precise since the function could conceivably agree almost everywhere with a continuous function, but more elaborate constructions along this line will work].

2) If you mean "Given any $\ell^2$ sequence $g(n)$ (for $n \in E$), there exists a continuous function $f$ on $\mathbb{T}$ such that $\hat{f}(n)=g(n)$ for $n \in E$, but may be arbitrary for $ n \notin E$," then this is equivalent to $E$ being a $\Lambda(2)$ set.

See W. Rudin's paper Trigonometric Series with Gaps, (in particular section 5.3).

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.